Shaping the field of Cardiovascular Fluid Mechanics: The 40th Anniversary of Ajit Yoganathan’s Research Laboratory

Pekkan, Kerem; Oshinski, John N.

doi:10.1007/s13239-021-00576-1

Cited by 1 publication

(1 citation statement)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ey applied eight-level metrics, including obtained high prediction accuracy [21]. Using decision trees and ensemble learning technology, the researchers created a dynamic software defect prediction model, the model uses the bagging ensemble learning method based on decision tree technology in the inner layer, a stochastic deep forest model is constructed, and random sampling and stacking methods are used in the outer layer to train these random forest models, and a certain prediction effect is achieved [22]. e investigators used logistic regression to estimate software faults on 37 indicators of antenna configuration software, identify static software metrics and the amount of software flaws, and find a correlation between the two [23,24].…”

Section: Literature Reviewmentioning

confidence: 99%

Defect Prediction Technology in Software Engineering Based on Convolutional Neural Network

Liu

Sanober

Zamani

et al. 2022

Security and Communication Networks

View full text Add to dashboard Cite

Software defect prediction has become a significant study path in the field of software engineering in order to increase software reliability. Program defect predictions are being used to assist developers in identifying potential problems and optimizing testing resources to enhance program dependability. As a consequence of this strategy, the number of software defects may be predicted, and software testing resources are focused on the software modules with the most problems, allowing the defects to be addressed as soon as feasible. The author proposes a research method of defect prediction technology in software engineering based on convolutional neural network. Most of the existing defect prediction methods are based on the number of lines of code, module dependencies, stack reference depth, and other artificially extracted software features for defect prediction. Such methods do not take into account the underlying semantic features in software source code, which may lead to unsatisfactory prediction results. The author uses a convolutional neural network to mine the semantic features implicit in the source code and use it in the task of software defect prediction. Empirical studies were conducted on 5 software projects on the PROMISE dataset and using the six evaluation indicators of Recall, F1, MCC, pf, gm, and AUC to verify and analyze the experimental results showing that the AUC values of the items varied from 0.65 to 0.86. Obviously, software defect prediction experimental results obtained using convolutional neural networks are still ideal. Defect prediction model in software engineering based on convolutional neural network has high prediction accuracy.

show abstract

Section: Literature Reviewmentioning

confidence: 99%